Welcome to Anote

Anote is an applied AI research company in New York City. Anote builds AI technology to provide high quality datasets and evaluations for leading enterprises, federal clients, and model providers.

Anote Core Products

Anote's core products operate to provide the full stack AI / ML lifecycle, from data curation, data annotation, model training, model inference, model evaluation, on premise deployment, and LLM integration.

The company develops Human-Centered AI designed to make AI more reliable, explainable, and effective for real-world use. Here are some of the areas that we focus on:

• Data curation and dataset creation
• Data annotation and human feedback pipelines
• Synthetic data generation
• Model training and fine-tuning
• Model inference and evaluation
• Retrieval-Augmented Generation (RAG) systems
• Multi-agent AI frameworks
• Private AI assistants and on-premise deployments

Research Publications

Anote conducts research in human-centered AI, machine learning optimization, and retrieval-augmented generation systems.
The following publications describe some of the core research contributions from the Anote team.

Improving Classification Performance with Human Feedback

Authors: Chung, E., Zhang, L., Jijo, K., Clifford, T., & Vidra, N.
Year: 2024

This paper introduces a framework for improving classification performance using human-in-the-loop learning.
By labeling a small subset of examples, the system can automatically infer labels for the remaining data while maintaining high accuracy.

Paper:
https://arxiv.org/abs/2401.09555

Enhancing Large Language Model Performance to Answer Questions and Extract Information More Accurately

Authors: Jijo, K., Setty, S., Chung, E., Vidra, N., & Clifford, T.
Year: 2024

This work explores techniques for improving the performance of large language models in question answering and information extraction tasks.
The research focuses on methods that combine structured prompts, evaluation workflows, and model optimization strategies.

Paper:
https://arxiv.org/abs/2402.01722

Improving Retrieval for RAG-Based Question Answering Models on Financial Documents

Authors: Setty, S., Thakkar, H., Lee, A., Chung, E., & Vidra, N.
Year: 2024

This paper investigates improvements to retrieval systems used in Retrieval-Augmented Generation (RAG) pipelines.
The research focuses on financial document analysis and demonstrates methods for improving the accuracy of question answering systems operating on complex financial datasets.

Paper:
https://arxiv.org/abs/2404.07221

Resources

Educational Repositories

These repositories support AI education initiatives, fellowships, and research training programs.

• https://github.com/anote-ai/btt-anote1a
• https://github.com/anote-ai/btt-anote1b
• https://github.com/anote-ai/btt-anote2a
• https://github.com/anote-ai/btt-anote2b
• https://github.com/anote-ai/BTT-Anote-1A-2024
• https://github.com/anote-ai/BTT-Anote-1B-2024
• https://github.com/anote-ai/BTT-Anote-1C-2024

Our Mission

Anote's mission is to make AI more accessible. To bridge the gap between the potential of AI models, and the every day tasks people care about. Anote is founded by Natan Vidra. For research collaborations, partnerships, or enterprise deployments, contact nvidra@anote.ai